NPI Parameters Released

Big news from the NCAA as they released the NCAA Power Index (NPI) parameters. I will list them below and then discuss the impacts. I’ll also link to the two NCAA documents that describe what they are doing.

The big parameter is “Win%/SOS”. The committee decided to go with 20/80. The big thing here is that Strength of Schedule is not a part of NPI at all. The NCAA should really stop using the term. This is the hardest point to get across to coaches when discussing NPI. SOS is not a thing. It’s not. Stop. What this parameter is saying is how much of my opponent’s NPI will count when I play them (win or lose). Your opponent’s NPI is not SOS. It’s NPI. You know how your 3^rd grade teacher told you not to use the word you are defining in the definition? That’s the basis of NPI.

* NOTE * I corrected a mistake above about 20/80. Thank you to one very sharp-eyed reader. I had it backwards and then complained that the NCAA had it backwards originally. Much funnier but entirely wrong on my part.

The next parameter is Home/Away bonus. The committee decided NOT to have any bonus. All matches, home, away or neutral will count the same. Two coaches mentioned to me that they are happy with this.

The Quality Win Bonus (QWB) was set at 55.5. I really read this as 0.555 because that’s how NPI is displayed in our simulated program. What this parameter means is that a winning team will get a bonus if they beat a team who has an NPI higher than 0.555. The further away (higher) an opponent’s NPI is, the bigger the bonus the winning team will get.  When my Latvian Hackers ran the 2023 season, they came up with 100 teams that had an NPI of 0.555 or higher. (Note – This was not simulated using the 10-match minimum, which we’ll discuss shortly.)

How much of a bonus is determined by the next parameter: QWB Multiplier. This is set at 0.6 and will be used to calculate how much of an NPI bonus a winning team will get when beating a team with an NPI of 0.555 or higher. (I assume in our program, the multiplier would be set to 0.06.)

The committee decided NOT to award a bonus (or penalty) for a 5-set match. A 25-0, 25-0, 25-0 loss is the same as 25-23,25-23, 23-25, 23-25, 13-15 loss. I find this interesting more than anything. Not sure what I would have done here. I don’t like the thought of penalizing a team for winning a 5-set match, but I do like rewarding teams for playing a tough opponent evenly.

The final parameter caught me a bit off guard. (Trust me, we will be looking at which Latvian is at fault for not letting me know sooner. It was probably Serge. I’m at my wit’s end with him.) This is “Minimum Wins” and it’s set at 10. In my past articles I’ve mentioned how playing poor teams doesn’t hurt. By ensuring that a team’s NPI is calculated by at least 10 wins, it sort of changes this dynamic a bit. Teams now must ensure that they have 10 quality matches they can win on the schedule before scheduling softer teams. Looking at our NPI numbers from 2023, I would categorize around 215 teams that would have helped an opponent if they won the match. That’s more than half of the teams in DIII volleyball. So, this is a good change by the committee but it really just eliminates some of the crazier paths towards getting an at-large bid by a team that tries to trick the formula.

The NCAA official NPI Report can be found here.

The second document from the NCAA is the FAQ and I’ll go through this briefly to highlight areas that are just plain wrong or interesting. You can find the entire FAQ, here.

The first and third (and I’m sure other) questions highlight how record and SOS are still used. As I mentioned above, SOS is not used at all. Yes, a match NPI is determined by who you play (their NPI) but the strength of a team’s schedule is not entirely factored in because wins against teams that lower your NPI are discarded. In other words, matches against poor teams that you beat are like they never happened. Until someone can show me how playing poor teams didn’t impact SOS in the past, I’ll continue to beat this drum that SOS is no longer used in selecting at-large teams.

Remember that common opponent and head-to-head are no longer used at all in NPI.

One of the benefits mentioned for NPI is that “This will greatly benefit stronger teams in weaker conferences.” This, at its crux, is what is wrong with NPI. This would be fine if these stronger teams played a strong out-of-conference schedule but they don’t have to in order to still get a good NPI. Maybe good enough for an at-large bid even, if they need one. Now that we’ve removed head-to-head, common opponent, ranked wins and SOS from the criteria, what safeguards do we have to ensure that these types of teams don’t crash the at-large bids?

There still will be Regional Advisory Committees (RAC) but one of their most important reasons for existing (ranking) will be gone. In fact, the FAQ says that regional rankings will still be published but it will be based off of NPI. I’d hope the entire list of teams is published DIII wide. If not, this is just stupid because it hides other teams in tough regions.

The parameters noted above can not be changed for up to 2-years.

The parameters were tested against the last three volleyball seasons. The FAQ says that only slight variances were noted. We need to rerun our numbers with the 10-match minimum but we saw 5 teams swap out in 2023. This was the most of any year we ran with our formula (which attempts to mimic NPI).

Some coaches are still confused by this but the FAX clearly states that this new process starts NOW.

In talking to a coach about this today, he indicated that each sport will have their own parameters. I had heard at one time that all sports had to share the same parameters. Thankfully, I was mistaken as it made no sense.

I should note that the FAQ does list the Win%/SOS parameter as 20/80 in their example so they got the ordering right there, at least.

“Strength of Schedule: SOS is simply an average of a team’s opponent’s NPI.” NO. IT. IS. NOT! They redefined what SOS is so it doesn’t scare people.

“Wait, SOS is no longer used? We can’t do that!”

“No. Don’t worry. We still use SOS.”

Heck they could have added that they also use unicorn tears in the formula for as much sense as it makes.

SOS (to the NPI formula) is your opponent’s NPI excluding wins against teams that are too lousy to consider. There. I fixed it.

They do have an examples of how QWB and the multiplier are handled. Basically, the NPI difference from the team you beat to the QWB of 0.555 is used. Let’s say the team you beat had an NPI of 0.575 so the difference is 0.020. You would then multiply that against the multiplier of 0.6 to get your bonus. I will point out that our formula uses percentage but it’s possible the real NPI will move some decimal points over to make it more readable. Not important as long as all of the decimal points get moved over together.

Conclusion

The absolute lowest (on “SOS”) the committee could have gone was 20/80. I think 15/85 would have been better but I also think there is an inherent flaw in the formula that hurts teams that play tougher schedules. So, I’m happy they didn’t go with 25/75 or lower. Still, it’s possible that undeserving teams (as defined by the old selection criteria and your eyeballs) will get at-large bids over much better teams. My only hope is that 10-match minimum fixes some of this. As I type, my Latvian Hackers are trying to run our formula with the new parameters to see how things would have turned out so stay tuned for that article.

I’m fine with the other parameters although it really eliminates any art (shout out to Ricky Nelson) from the selection at all. No head-to-head is considered. You may be the better team on the court where it counts but it no longer counts because NPI is the only one allowed to count. Common opponent was always a bit of nightmare to consider but it gave us a picture of how teams did against the same opponent. Ranked Results was never done quite right but when it was, you could consider how teams did against the best.

Bottom line, the robot overlords have taken over DIII Volleyball selection. We all thought they would have started somewhere more powerful like the White House but it appears their plan is more sinister. We’ll have to wait to see how it all turns out and, in truth, this will only impact 4 to 10 teams at the end of the year when we see how the bubble teams are handled. Then, potentially, the resistance will gain in strength and the fight will begin.